Netgate Discussion Forum
    • Categories
    • Recent
    • Tags
    • Popular
    • Users
    • Search
    • Register
    • Login

    PFSENSE breaks VPN after Upgarde

    Scheduled Pinned Locked Moved IPsec
    7 Posts 3 Posters 1.0k Views 3 Watching
    Loading More Posts
    • Oldest to Newest
    • Newest to Oldest
    • Most Votes
    Reply
    • Reply as topic
    Log in to reply
    This topic has been deleted. Only users with topic management privileges can see it.
    • F Offline
      filoips
      last edited by

      I got several VPN IPSEC towards Siphos XF Firewall.
      They worked like a charm until I upgraded to 21.02.2-RELEASE and subsequent patch p1

      PFsense stop processing IPsec, if it found TCP connection over VPN , packet will be dropped after 20 seconds.

      just to be more clear :
      2 site

      • A with workstation WA and WB ( managed by Sophos XG )
      • B with workstation WC (Managed By PFSENSE )

      VPN always Up, no problem for rekey and so on ..
      SPI aligned and so on ...

      WA -> pinging WC
      and
      WB -> pinging WC

      WB -> start TCP session to WC ( for example ssh )

      after about 20 seconds ( just login to ssh and get remote prompt )

      WA -> Pinging WC like a charm
      WB -> NOT able to ping WC anymore, nor othe desktop on site B
      All traffic from WB is neither visible through ENC interface

      I can reproduce it at anytime

      very strange

      If I restart VPN , traffic start to flow again until from WB I start a tcp connection towards WC

      I
      any help would be appriciate

      PS: Open a ticket to Sophos Support, they didn't found any problem
      PS1: dump traiffc on pfsense fro proto 50 on wan interface I got ESP packet with right SPI but that packet ( if it were processed ) didn't appera on enc0 interface not, on any other

      1 Reply Last reply Reply Quote 0
      • S Offline
        sgnoc
        last edited by

        @filoips I'm wondering if I am having the same problem you're describing. I maintain an IPSec tunnel to another endpoint for file sharing mostly. I don't use it frequently, so I can't say exactly when this problem occurred, but I haven't been able to figure it out quite yet.

        I have a standard IPSec Tunnel from networks A and B.
        Network B is a network I don't have management control over, so I only have knowledge of my end.

        On my network, I can restart the tunnel and it stays connected with no problems. Both the P1 and P2 seem to be connecting and maintaining the connections properly.

        I can ping the server on Network B from any systems on Network A and get responses. As soon as I attempt a CIFS connection from Network A to the file server on Network B, I can sometimes get enough data through to load the directory structure before the link becomes unresponsive.

        In some cases I can't even get a directory listing from the CIFS connection, in others I can get the main directory listing before it becomes unresponsive, and rarely I can get a sub directory listing and then it will become unresponsive. I've not been able to do more than that.

        The tunnel still shows connected, but CIFS connections timeout and I get no replies from pings with systems on Network A to Network B.

        I've tried CIFS connections from multiple computers on Network A with the same result.

        I can either reconnect the P2 entry to reset everything, or if I wait an undetermined amount of time, I can come back and get ping responses from the server on Network B.

        Would you think this sounds similar to the connectivity issues you are having? I don't have a way to check for other TCP connections other than the CIFS connection going out. This link used to work without issues, though.

        If not, I'll start up a new topic.

        1 Reply Last reply Reply Quote 1
        • S Offline
          sgnoc
          last edited by

          @filoips I continued doing some research and found a (temporary) solution. Apparently something broke with the hardware encryption in the latest release.

          Here is a good read: https://forum.netgate.com/topic/161354/ipsec-packet-loss-routing-issue-with-21-02-release

          I tried 2 options, first was to change hardware encryption to use QAT from the AES-NI. That seemed to get my connection a little more stable. I was able to go to a few more directories before it became unresponsive.

          Then I changed to none to completely disable hardware encryption. Rebooted pfsense to unload the module, and now I have a stable and working connection again with no loss in connectivity.

          Hopefully that helps and hopefully the next release will fix whatever was broken.

          F 2 Replies Last reply Reply Quote 1
          • F Offline
            filoips @sgnoc
            last edited by

            @sgnoc I think same issue, very strange problem, today I'm goign to change Algo and then try to disable hardware encryption.

            I know new firmware is planned on june ( 21.05 ), but from release notes there is any reference for this issue

            1 Reply Last reply Reply Quote 0
            • F Offline
              filoips @sgnoc
              last edited by

              @sgnoc I moved to OLD Style Algo
              3des-md5-dh2
              and vpn start to be stable
              I know they are week algo ... :-(

              1 Reply Last reply Reply Quote 0
              • N Offline
                NOCling
                last edited by

                Go for 21.05 the Hardware crypto Problems are solved.
                https://forum.netgate.com/topic/164159/pfsense-plus-21-05-release-now-available

                Netgate 6100 & Netgate 2100

                1 Reply Last reply Reply Quote 0
                • S Offline
                  sgnoc
                  last edited by

                  Thanks for the heads up on the update @NOCling

                  Just got the upgrade in and reconfigured the AES-NI and everything seems to be working again. The AES-NI is confirmed to be back on and my IPSec tunnles are working at optimum speeds again. Seems to be fixed for my purposes anyway.

                  Hope it resolves any issues for everyone else!

                  1 Reply Last reply Reply Quote 0
                  • First post
                    Last post
                  Copyright 2025 Rubicon Communications LLC (Netgate). All rights reserved.